For the longest time, the belief was that building a foundational AI model required massive computing power (specialized chips like Nvidia’s) and billions of dollars in investment. The dominance of US tech giants like OpenAI, Microsoft, and Google reinforced this idea.

But DeepSeek, an AI model from China, changed the narrative. Despite restrictions on high-end AI chips due to US export controls, DeepSeek delivered high-quality results at a fraction of the cost—reportedly just $6 million. This is a huge contrast to US companies spending billions on similar AI models.

Why This Matters

1. Breaking Assumptions

i) It challenges the idea that AI requires unlimited resources.

ii) If AI can be trained efficiently at lower costs, Nvidia and other chipmakers could take a hit.

2. The Open-Source Factor

i) Unlike US firms that guard their models, DeepSeek is open-source, allowing global contributions.

ii) This shifts AI development from a few powerful corporations to a broader community.

3. Questions Around Cost

i) The $6M claim only covers chip rental costs, not research, salaries, or prior experiments.

ii) Experts are still verifying whether such a low-cost AI model is truly possible.

Why Didn’t India Build Its Own DeepSeek?

A key reason is market dynamics. Unlike China, which restricts US AI companies, India has an open market.

US AI models dominate in India, making it tough for local AI startups to compete.

In China, local companies get time to improve and scale up without foreign competition.

The Bigger Picture

If DeepSeek’s model is truly efficient, AI development costs could drop dramatically, leading to:

i) More affordable AI solutions.

ii) A rethink of AI investments by US giants.

iii) Possible disruptions in Nvidia’s dominance.

Final Thought: DeepSeek’s rise is a wake-up call for the AI industry. Whether it’s truly a low-cost revolution or just clever marketing, its impact is undeniable.